Chronological Thread 

< Chronological >      < Thread >

---

Hi Mike,
I do not have an SS7 background. However, I will try to answer your
questions
and hope they will be helpful to you. You might be able to find better
answer by
posting the question to SIGTRAN group.
First thing to note is that the concept of SCTP stream is different from
the
underlying number of links. Even though you have 16 physical links in an
linkset, you can have more than 16 streams or less than 16 streams. Having
said that, if you negotiate 16 streams between SG and MGC, mapping of
messages from stream id to SLS and vice versa (For messages received
from MGC to SG and vice versa) would be much easier than having less than
16 streams or more than 16 streams.
So I would say that number of streams per association should be chosen
depending on the nature of application. In case of IUA, you could negotiate
number of streams equal to number of D channels supported by the SG. If
SG is capable of supporting 4K ISDN lines (Either BRI/PRI), both SG and
MGC could negotiate 4096 streams. In this case if you use only default
stream 0, then the risk of head-of-line blocking when transferring
messages between SG and MGC are high.

If you post your message

-thanks,
Umesh

> -----Original Message-----
> From:
> discussion-admin AT sctp.de
>
> [mailto:discussion-admin AT sctp.de]On
> Behalf Of Michael Mueller/bhu5nji
> Sent: Tuesday, December 18, 2001 9:40 AM
> To:
> discussion AT sctp.de
> Subject: Re: [SCTP-Discussion] Streams and D-Channels
>
>
> Umesh,
>
> Thanks for clearing that up.
>
> This discussion has raised questions in my mind about the general
> use of streams within an association. My background is in large
> part from SS7 so that flavors my questions.
>
> In reading http://www.sctp.de/papers/drcn2001.pdf, Section 5
> describes the stream concepts as being key to avoiding head-of-line
> blocking. Does this mean that in designing an SS7 linkset to SCTP
> gateway, an attempt should be made to negotiate 32 streams to
> accommodate 32 SLS for an ANSI combined linkset (or 16 streams for
> 16 SLS for an ITU linkset)? Or should each physical link in a
> linkset be mapped to a separate stream?
>
> Why are streams negotiated? What issues lie behind the
> negotiations? Why for instance, would a "server" end of an
> association reject the "client" end's request for n streams?
>
> I am also curious about why sctp sending has an optional field to
> specify the stream for transmission. My SS7 sensibilities are
> confused by the optional nature of this piece of information. What
> happens when two SCTP entities are connected where one uses multiple
> streams and the other uses the default stream 0 only? Is the
> outcome that the performance of the entity using the default stream
> only is at risk of being comparatively poorer because the risk of
> head-of-line blocking is increased?
>
> Mike
>
> ----- Original Message -----
> From:
> <umesh AT cosystems.com>
> To:
> <discussion AT sctp.de>
> Sent: Sunday, December 16, 2001 10:21 PM
> Subject: RE: [SCTP-Discussion] Streams and D-Channels
>
>
> : Hi Mike,
> : You are right. The QPTM (Q.921/Q.931 boundary primitives
> Transport)
> : contain IUA header
> : in addition to common header. The IUA header has a DLCI (SAPI+TEI)
> field
> : which identifies
> : the D channel corresponding to this message.
> :
> : Umesh.
> :
> : > -----Original Message-----
> : > From:
> discussion-admin AT sctp.de
> [mailto:discussion-admin AT sctp.de]On
> : > Behalf Of Michael Mueller/bhu5nji
> : > Sent: Saturday, December 15, 2001 3:29 PM
> : > To:
> discussion AT sctp.de
> : > Subject: [SCTP-Discussion] Streams and D-Channels
> : >
> : >
> : > :
> : > : And, Yes u can send on a single StreamID mutiple D-Channel
> : > Signaling mesgs.
> : >
> : > How would you know which D-Channel a signal is from or going to?
> Is
> : > there something in the IUA header that differentiates the
> channels?
> : >
> : > MIke
> : > :
> : >
> : >
> : > _______________________________________________
> : > discussion mailing list
> : >
> discussion AT sctp.de
> : > http://www.sctp.de/mailman/listinfo/discussion
> : >
> : >
> :
> : _______________________________________________
> : discussion mailing list
> :
> discussion AT sctp.de
> : http://www.sctp.de/mailman/listinfo/discussion
>
> _______________________________________________
> discussion mailing list
> discussion AT sctp.de
> http://www.sctp.de/mailman/listinfo/discussion
>
>

---

• [SCTP-Discussion] RE: discussion digest, Vol 1 #198 - 3 msgs, hari, 12/14/2001
  • [SCTP-Discussion] Streams and D-Channels, Michael Mueller/bhu5nji, 12/14/2001
    • RE: [SCTP-Discussion] Streams and D-Channels, umesh, 12/17/2001
      • Re: [SCTP-Discussion] Streams and D-Channels, Michael Mueller/bhu5nji, 12/17/2001
        • RE: [SCTP-Discussion] Streams and D-Channels, umesh, 12/18/2001

---

Archive powered by MHonArc 2.6.19+.